[asvs] Re: Visual Semantics
- From: "Dave Reynolds" <dkreynolds@xxxxxxxxxxxx>
- To: <asvs@xxxxxxxxxxxxx>
- Date: Sun, 31 Oct 2004 19:08:24 -0000
I agree that the second method is preferable. This could be enhanced by
allowing the user to gradually build a template which as well as being
stored mentally would be stored on the user's machine. This in some respects
limits the user, because the template will not follow him from system to
system. However, it will serve as a method of firmly establishing a vi
user's interpretation of objects shapes and their relative positions.
Dave.
----- Original Message -----
From: "Will Pearson" <will-pearson@xxxxxxxxxxxxx>
To: <asvs@xxxxxxxxxxxxx>
Sent: Sunday, October 31, 2004 6:41 PM
Subject: [asvs] Visual Semantics
> Hi;
> A couple of weeks ago, there was some discussion on alternate forms of
> synthetic vision in which semantics came up. I've been giving this some
> thought, as it's one area of research in which I've done some work with
> access to diagrams. Here's my basic thoughts.
>
> I think it's possible to produce some form of artificial vision through
> semantics, but the big question is, do we allow the user to establish the
> meaning for themselves or do we inform the user of the meaning associated
> with a shape?
>
> The same shape can represent different objects within different contexts,
> and the same shape can take on different meanings within different
> contexts. This is a problem if the meaning were to be given to the user.
> A potential system would need to know the context in which the object was
> placed, and the object's meaning within that context.
>
> To some extent, context can be derived from the shapes and objects around
> the one for which interest has been placed. This can help to establish
> context, but may not completely establish the context for the shape, and
> thus make automatically establishing a context a bit more difficult.
>
> Another way is to allow the user to establish the meaning for themselves.
> This more closely approximates communication where psychology,
> communications theory and semiotics all suggest we, as humans, separate
> the stimuli we receive from the meaning we associate with it. In doing
> this, we allow for factors other than the pure sensory stimuli to
> influence the meaning associated with what we're experiencing. Therefore,
> a potential system could purely indicate shapes and their attributes,
> along with the spatial relationships between other shapes, and allow a
> user to associate meaning with the information. This will likely increase
> the TLX for the user, but may offer a more accurate mechanism for
> associating meaning with shapes than a purely automated system.
>
> Neither of these approaches deals with visual aesthetics. However, the
> primary use for sight is a functional one, in which we require the meaning
> conveyed by what we see, not particularly how something looks. So, by
> conveying meaning and spatial relationships, it should be possible to gain
> this functional aspect of vision.
>
> Will
>
----------------------------------------
My Inbox is protected by SPAMfighter
3519 spam mails have been blocked so far.
Download free www.spamfighter.com today!
Other related posts:
